Processing an list vectors at once

[mllobera]

I could not find this information so I thought I would ask here.

Is there a way to be able to process an array or list of vectors all at once?

So something like,

lst = [Vector3([1,2,3]), ..., Vector3([7,8,9])]
rot_lst = lst * Matrix44.from_x_rotation(radians(90))

I have seen that there is a function matrix44.apply_to_vectors() but I have been unable to get it to work with either a numpy array or a list (also I cannot tell when to use Matrix44 or matrix44? it seems that there are two different set of methods???)

[adamlwgriffiths]

Support for mass transforms is inconsistent and depends on the function.
The intention was always to add support for it, but I never got around to it.

A trivial (and slow) way to do it would be detect lists of lists / arrays of arrays, and iterate over them at the function level using a decorator.

[mllobera]

Thanks for the quick response. It is good to know.

…

On Fri, Nov 15, 2024 at 8:25 AM Adam Griffiths ***@***.***> wrote: Support for mass transforms is inconsistent and depends on the function. The intention was always to add support for it, but I never got around to it. A trivial (and slow) way to do it would be detect lists of lists / arrays of arrays, and iterate over them at the function level using a decorator. — Reply to this email directly, view it on GitHub <https://urldefense.com/v3/__https://github.com/adamlwgriffiths/Pyrr/issues/127*issuecomment-2479357055__;Iw!!K-Hz7m0Vt54!lk7dQgC_MeXjflyB0QsPaFGTWgAS9AE_QYGtE6t03OFHjwGpUaB39m8UzAnOR21_WBVP37YCP-gn64H7eIithnA$>, or unsubscribe <https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AANCY3VLY5HMQDZPZ2CNKNL2AYN53AVCNFSM6AAAAABR2DU7GGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZZGM2TOMBVGU__;!!K-Hz7m0Vt54!lk7dQgC_MeXjflyB0QsPaFGTWgAS9AE_QYGtE6t03OFHjwGpUaB39m8UzAnOR21_WBVP37YCP-gn64H7fI3Ueo8$> . You are receiving this because you authored the thread.Message ID: ***@***.***>

--

___________________ Marcos Llobera Assoc. Prof. in Anthropology Digital Archaeology Research Lab (DigAR Lab <https://www.digarlab.uw.edu/>) University of Washington pronouns: *he/his/him*

[mllobera]

The following example is a hack but it works to do mass transformation (the price you pay is that you revert back to numpy).

S = Matrix44.from_scale([0.5, -0.5, 1])
R = Matrix44.from_z_rotation(radians(90))
T = Matrix44.from_translation([-100, -100, 0])
t = S * R * T

vecs = np.array([[100.0, 100.0, 0, 1],
                           [200.0, 200.0, 0, 1]])

trans = vecs * np.matrix(t)
trans = np.array(trans)

result

array([[ 0., 0., 0., 1.],
[50., 50., 0., 1.]])

Not ideal but...

[adamlwgriffiths]

I haven't tested it, but that should work as is, since the procedural and oo interfaces are just np.arrays. Possibly some of the extra stuff added will cause problems.

Vector3 -> BaseVector3 -> BaseVector -> BaseObject

Alternatively you can just use the procedural API which is literally just np.array.

[mllobera]

Thanks Adam. I was surprise to see that Matrix44 did not behave as a numpy matrix. By just perusing very quickly through your code, this is explained because , as you point out, matrix inherits from an numpy array and not a numpy matrix. Changing this would represent a huge endeavor but I am wondering whether it would provide a slick way towards providing a way for mass transformations. Just some thoughts...